Statistical methodology research

We have particular research strength in the development of Bayesian methods, modern regression methodology, semiparametric methods, and survey design and analysis.

Bayesian methods

We develop Bayesian methodology for posterior sampling and model selection, and tools for use in several other fields of research including astrostatistics, ecology, econometrics, engineering, fisheries, forensics, genomics and medicine.

Modern regression

We develop novel ways to make inference about variables of interest using information gathered on other variables. Examples include extensions of generalized linear models and smoothing techniques, and tools for analysis of longitudinal data, high-dimensional data, and spatial and temporal data.

Semiparametric and nonparametric methods

We develop more reliable semiparametric estimators, methods for smooth estimates of densities and hazard functions, and mixture models.

Complex surveys and design

Analysing data from complex surveys can be misleading if treated as a simple random sample. Our research includes finding new ways of designing and analysing data from many types of studies, market research, panel surveys, multi-level surveys, and industrial and laboratory experiments.